Triple Stores

What is a triple store? A database for RDF triples Can ingest RDF in a variety of formats Supports a query language SPARQL is the W3C recommendation Other RDF query languages exist (e.g., RDQL) Might or might not do inferencing Triples stored in memory in a persistent backend Persistence provided by a relational DBMS (e.g., mySQL) or a custom DB for efficiency.

Architectures Can be divided into several categories: In-memory, Native store, Non-native store In memory: RDF Graph is stored as triples in main memory Native store: Persistent storage systems with custom DBs, e.g.: JENA TDB, Sesame Native, Virtuoso, AllegroGraph, Oracle 11g Non-Native store: Persistent storage systems set-up to run on third party DBs, e.g., Jena SDB using mysql or postgres RDF stores can be divided into 3 broad categories based upon how they are implemented. These are not precisely defined terms with rigid boundaries but something that can serve as rough guide to what a triple store is. Native Stores: Provide support for transactions, own SQL compiler and generally their own procedure language. Will explain in detail what I mean by this term in the later slides.

Architecture trade-offs In memory is fastest, obviously, but load time has to be factored in Native stores are fast, scalable, and popular now Non-native stores may be better if you have a lot of updates and/or need good concurrency control See the W3C page on large triple stores for some data on scaling for many stores

Large triple stores in 2018 http://www.w3.org/wiki/LargeTripleStores

Quads, Quints and Named Graphs Many triple stores support quads for named graphs A named graph is just an RDF with a URI name often called the context Such a triple store divides its data a default graph and zero or more additional named graphs SPARQL has support for named graphs De facto standards exist for representing quad data, e.g., n-quads and TriG (a turtle/N3 variant) AllegroGraph stores quints (S,P,O,C,ID), the ID can be used to attach metadata to a triple

Support for Reasoning Most triple stores don’t do much (or any) reasoning and use a simple model: You do the reasoning to materialize all of the triples you want, which you then load into the store Triple store provides query and update APIs, access control, SPARQL interface, efficient indexing, etc. Some do support reasoning, e.g., Jena has a native rules engine and an API for external reasoners (e.g., Pellet, Fact++) Sesame has a native RDFS reasoner Stardog supports OWL DL reasoning via query expansion

Example: Jena Framework An open software Java system originally dev-eloped by HP (2002-2009) Moved to Apache when HP Labs discontinued its Semantic Web research program https://jena.apache.org/ Using the TDB native store, it can easily handle ~2B triples Good tutorials and documentation Has internal reasoners and can work with DIG compliant reasoners such as Pellet Supports a Native API and SPARQL via Fuseki

Example: Sesame Sesame is an open source RDF framework with support for RDFS inferencing and querying http://www.openrdf.org/ Implemented in Java Query languages: SeRQL, RQL, RDQL and SPARQL Triples can be stored in memory, on disk, or in a RDBMS Has a native RDFS reasoner Easy to setup & use, but tops out at ~70M triples

Example: Stardog http://stardog.com/ by Clark and Parsia Pure Java RDF database (“quad store”) Lightweight and very fast for in-memory use Reasoning support via Pellet for OWL DL and query rewriting for OWL 2 QL, EL & RL Command line interface and JAVA API Commercial, but has a free version good for modest projects ~50B triples on $10K server with 256G ram and 32 cores

Performance Much work on benchmarking of triple stores There are several standard benchmark sets Two key things are measured include Time to load and index triples Time to answer various kinds of SPARQL queries The Berlin SPARQL Benchmarks evaluated 4store, BigData, BigOwlim, Jena TDB and Virtuoso in 2011 with 100M and 200M datasets. The numbers are “query mixes per hour”, so bigger is better

Load Time

Queries per hour

Summary A triple store is an essential component of any system using RDF There are a number of good ones available, both open sourced and commercial Developing triple stores for large-scale parallel systems is still a research topic